The proposed research is designed to increase our knowledge of the basic processes of reading. The studies will examine comprehension of textual materials and memory structures resulting from the comprehension process. A major aspect of comprehension is the process of inference. We propose a simple model of inference and use it as a framework to design experiments that investigate under what conditions to-be-inferred information in long-term memory is activated, under what conditions it is connected to explicitly stated information, and under what conditions it is encoded with the explicitly stated information in the long-term memory representation of the text. We also propose several experiments to investigate in detail the organization in memory of textual information. The task employed in many of the proposed studies is item recognition. Subjects decide whether an item (word or proposition) was in a text (sentence or paragraph) they read. The dependent measure is the time required for a correct response. Response time will be faster, for example, if the item tested is active, or if the immediately preceding test item was connected (or related) to this test item in the memory representation of the text. One section of the proposal is devoted to examining the automatic and strategic components of this activation (or priming) process. Subjects in the experiments will be college undergraduates.